The gist: Natural language → production data pipelines. Built on Arrow, WebAssembly, and CRDTs for real-time collaboration. But here's what matters: A PM can type "analyze user cohort retention" and get a working pipeline. No DAGs, no YAML, no suffering.
Happy to answer technical questions. Yes, it handles backpressure. Yes, exactly-once semantics. Yes, you can drop down to code when needed. What data pipeline are you putting off building because it's too much hassle?